CN101605089A - BGP dynamic migration method and device - Google Patents

BGP dynamic migration method and device Download PDF

Info

Publication number
CN101605089A
CN101605089A CNA2008101147347A CN200810114734A CN101605089A CN 101605089 A CN101605089 A CN 101605089A CN A2008101147347 A CNA2008101147347 A CN A2008101147347A CN 200810114734 A CN200810114734 A CN 200810114734A CN 101605089 A CN101605089 A CN 101605089A
Authority
CN
China
Prior art keywords
bgp
gateway protocol
message
protocol process
border gateway
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CNA2008101147347A
Other languages
Chinese (zh)
Other versions
CN101605089B (en
Inventor
陈双龙
涂伯颜
张立新
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to CN2008101147347A priority Critical patent/CN101605089B/en
Publication of CN101605089A publication Critical patent/CN101605089A/en
Application granted granted Critical
Publication of CN101605089B publication Critical patent/CN101605089B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The embodiment of the invention discloses a kind of BGP dynamic migration method and device, this method comprises: create virtual neighbours according to the described dynamic migration order that receives and be connected with backup TCP; Obtain the routing iinformation of transferring neighbor in the configuration information on the Border Gateway Protocol process of source and the described transferring neighbor first inlet routing table on the Border Gateway Protocol process of source, it is placed in the middle to store described configuration information into described virtual neighboring, the routing iinformation of the described first inlet routing table is stored in the second inlet routing table of transferring neighbor on object boundary gateway protocol process of establishment, and carry out routing according to the routing iinformation that stores in the described second inlet routing table; The message data that TCP on the Border Gateway Protocol process of described source is connected corresponding described transferring neighbor receives the backup TCP connection.The embodiment of the invention can realize the dynamic migration of Peer between distributed BGP, guarantees session and professional not disconnection in the transition process.

Description

BGP dynamic migration method and device
Technical field
The present invention relates to moving communicating field, relate in particular to a kind of BGP dynamic migration method and device.
Background technology
The transmission of Internet route at present mainly relies on Border Gateway Protocol (Border GatewayProtocol, abbreviation BGP) Routing Protocol is finished, along with the IP technology rapid development, whole world Internet number of users is more and more, Internet route quantity is also in sustainable growth, and this just brings new challenge to bgp protocol.Present international gateway networking has following characteristics:
(1) neighbours (Peer) number of external BGP (External Border Gateway Protocol is called for short EBGP) is more, and scene commonly used needs 16, and extreme scenes can reach 100-200;
(2) the route amount is big, and at present the number of prefixes of Internet is about 230,000, and grows with time, and this quantity also can continue to increase;
(3) the route characteristic distributions is remarkable, and promptly the different neighbours' of same prefix route umber is extremely many;
(4) continue to exist route flapping, according to the related data introduction, average 6/s.
Handle a plurality of Peer in order to solve an above-mentioned CPU, the load that many routes are brought is excessive, the defective that pressure is overweight, BGP is carried out distributed design, Peer is distributed in the different bgp process, each BGP only handles and own relevant Peer like this, and the processing distribution of original centralized is come.Based on the distributed BGP of Peer (Peer-Distributed BGP is called for short PD-BGP) structural representation as shown in Figure 1, PD-BGP mainly finishes the reception of route and sends and handles.PD-BGP carries out the part routing to the route that the Peer that receives sends, only the local optimum route is sent to center BGP (Center BGP, be called for short CBGP), can make like this and only preserve those on the CBGP and may become preferred routing iinformation, reduced CBGP taking to internal memory; Information that PD-BGP sends Peer or the information that sends to Peer are carried out the filtration of ingress policer and egress policy, can reduce CBGP taking CPU like this.The CBGP major function is:
(1) receives the local optimum route that each PD-BGP sends and carry out full routing, select global optimum's route;
(2) generate forwarding-table item by global optimum's route, instruct and transmit;
(3) global optimum's route is sent to PD-BGP, so that issue Peer;
On function, CBGP has realized the basic function of all BGP, the ingress policer and the egress policy that comprise neighbours, Peer that still can normal configuration BGP on the CBGP, the flow chart of CBGP reception route as shown in Figure 2, in this flow process, Adj-RIB (Routing Information Base, be called for short RIB, route information table)-IN is the inlet routing table of BGP Peer, BGP can form an Adj-RIB-IN at each Peer, is used for depositing the routing iinformation of receiving from this Peer, and the routing iinformation of depositing is not through routing.Loc-RIB is the local routing table of BGP, and BGP carries out route selection to the routing iinformation that each Peer receives, preferred routing iinformation is stored among the Loc-RIB.Route among the Loc-RIB is used for forming forwarding information table (Forwarding InformationBase is called for short FIB), and this FIB is sent to Peer.Adj-RIB-OUT is the outlet routing table of BGP Peer, and BGP can send to corresponding Peer with preferred routing iinformation, and the route record that sends to Peer is in Adj-RIB-OUT.
Because the load of system is relevant with the run mode of system, such as route quantity, vibration, strategy etc., so it is unbalanced the load of some PD-BGP to occur, at this moment just certain Peer need be moved to the another one bgp process from a bgp process; Perhaps according to the development of business, certain bgp process can not have been carried the Peer of this BGP, at this moment needs some Peer of this BGP is moved in other bgp process, to reduce the load of this bgp process.
In addition, because professional development is dynamic, so the deployment in early stage possibly can't be satisfied the requirement of back autgmentability, at this moment just Peer need be moved in the another one bgp process from a bgp process dynamically, at this moment will accomplish to cause service disconnection because of dynamic migration.In addition for the Peer that reduces BGP switches the burden of bringing to network between process as far as possible, requiring in the handoff procedure of BGP Peer can not influence BGP Peer session under the prerequisite that does not cause service disconnection, and other business.The triggering of transition process can be a manual configuration, it also can be the load that system detects each bgp process automatically, such as taking CPU time, neighbours' number or the like, in case find that the load of certain bgp process is bigger, then trigger from this bgp process automatically, some BGP Peer are moved in the another one bgp process.This transition process is not subjected to the restriction of hardware device, can occur between each bgp process on the physical equipment plate.Also can occur on two blocks of physical equipment plates.
In order to solve the problems referred to above in the BGP Peer dynamic migration process, the BGP Peer that present a kind of method has adopted first deletion to move, reconfigure and be deployed to the scheme among other PD-BGP again, but this method has increased the workload of configuration, comparatively loaded down with trivial details, and delete the routing iinformation that the BGP Peer that needs migration also can delete this Peer, can cause losing of forwarding information, the normal business of influence.Present another kind of method has adopted BGP Peer session has been disconnected earlier, by Graceful Restart (GracefulRestart, be called for short GR) the not disconnection of technique guarantee business, and then the corresponding configuration dynamic migration of the Peer that will move is to the scheme of other bgp process, but this method causes the session disconnection, again the time of the company's of building time and data smoothing long, in time the response to network change in topology is very big to web influence; And require local terminal and opposite equip. must support the GR ability, to the equipment requirements height.
Summary of the invention
The embodiment of the invention provides a kind of BGP dynamic migration method and device, to solve the defectives such as information dropout, session and professional disconnection that occur in the prior art Peer transition process.
The embodiment of the invention provides a kind of BGP dynamic migration method, comprising:
Receive the dynamic migration order, and create virtual neighbours according to described dynamic migration order and be connected with backup TCP;
Obtain the routing iinformation of transferring neighbor in the configuration information on the Border Gateway Protocol process of source and the described transferring neighbor first inlet routing table on the Border Gateway Protocol process of source, it is placed in the middle to store described configuration information into described virtual neighboring, the routing iinformation of the described first inlet routing table is stored in the second inlet routing table of transferring neighbor on object boundary gateway protocol process of establishment, and carry out routing according to the routing iinformation that stores in the described second inlet routing table;
The message data that TCP on the Border Gateway Protocol process of described source is connected corresponding described transferring neighbor receives the backup TCP connection.
To achieve these goals, the embodiment of the invention also provides a kind of BGP dynamic migration device, comprising:
The migration preparation module is used for creating virtual neighbours according to the dynamic migration order that receives and is connected with backup TCP;
The route memory module, it is placed in the middle that the configuration information of transferring neighbor on the Border Gateway Protocol process of source that is used for obtaining stores described virtual neighboring into, routing iinformation in the first inlet routing table of transferring neighbor on the Border Gateway Protocol process of source that receives is stored in the second inlet routing table of transferring neighbor on object boundary gateway protocol process of establishment, and carry out routing according to the routing iinformation that stores in the described second inlet routing table;
The message receiver module is used for that the message data that the TCP on the Border Gateway Protocol process of described source connects corresponding described transferring neighbor is received backup TCP and connects.
The BGP dynamic migration method and the device of the embodiment of the invention, be connected with backup TCP by creating virtual Peer in the target bgp process, migration Peer all routing iinformations on the bgp process of source are all moved in the inlet information table of the corresponding migration of target bgp process Peer, and the message data of corresponding migration Peer in the TCP connection of source bgp process is all moved to backup TCP connect, make virtual Peer become the Peer of target bgp process.Because this Peer remains with source bgp process or target bgp process and is connected in transition process, therefore guaranteed normally carrying out of professional and session.
Description of drawings
Fig. 1 is the distributed BGP structural representation based on Peer;
Fig. 2 receives the flow chart of route for CBGP;
Fig. 3 is embodiment of the invention BGP dynamic migration method embodiment one flow chart;
Fig. 4 is embodiment of the invention BGP dynamic migration method embodiment two flow charts;
Fig. 5 is embodiment of the invention BGP dynamic migration method embodiment three flow charts;
Fig. 6 is embodiment of the invention BGP dynamic migration method embodiment four flow charts;
Fig. 7 is embodiment of the invention BGP dynamic migration device embodiment one structural representation;
Fig. 8 is embodiment of the invention BGP dynamic migration device embodiment two structural representations.
Embodiment
The embodiment of the invention is based on the distributed BGP of Peer, as shown in Figure 3, is embodiment of the invention BGP dynamic migration method embodiment one flow chart, specifically comprises:
010, receives the dynamic migration order, and create virtual Peer and backup transmission control protocol (Transmission Control Protocol is called for short TCP) connection according to this dynamic migration order;
020, obtain migration Peer at configuration information on the bgp process of source and migration Peer the routing iinformation of first inlet in the routing table on the bgp process of source, store the routing iinformation of the first inlet routing table migration Peer of establishment in the second inlet routing table of target bgp process, routing iinformation according to storage carries out routing, and the configuration information of migration Peer on the bgp process of source that receives stored among the virtual Peer;
030, the message data with corresponding migration Peer in the TCP connection of source bgp process receives the backup TCP connection.
The embodiment of the invention is connected with backup TCP by creating virtual Peer in the target bgp process, migration Peer is moved in the inlet information table of the corresponding migration of target bgp process Peer at all routing iinformations of source bgp process, and the message data of corresponding migration Peer in the TCP connection of source bgp process is all moved to backup TCP connect, make virtual Peer become the Peer of target bgp process.Because this Peer remains with source bgp process or target bgp process and is connected in transition process, therefore guaranteed normally carrying out of professional and session.
As shown in Figure 4, be embodiment of the invention BGP dynamic migration method embodiment two flow charts, because on an equipment, bgp process can comprise two class processes: CBGP process and PD-BGP process, bgp process is the general designation of CBGP process and PD-BGP process, and the migration of Peer can be carried out between two class processes, also can carry out between two PD-BGP processes.Present embodiment carries out dynamic migration with BGP Peer between different PD-BGP processes, concrete, moving to target P D-BGP process (present embodiment is designated as PD-BGP process 2) with migration Peer from source PD-BGP process (present embodiment is designated as PD-BGP process 1) is that example describes, and specifically comprises:
200, PD-BGP process 2 receives the dynamic migration order, creates virtual Peer and is connected with backup TCP;
In the present embodiment, the virtual Peer that PD-BGP process 2 is created is used to store the configuration information that moves Peer; The TCP connection that the backup TCP connection is used to store PD-BGP process 1 is to moving the message data of Peer, and this backup TCP connects not trigger this moment sets up new connection.
210, PD-BGP process 2 sends the ready message of migration to PD-BGP process 1, and creates the inlet routing table of migration Peer on PD-BGP process 2;
The inlet routing table that PD-BGP process 2 is created is designated as Adj-RIB-IN21;
220, PD-BGP process 1 receives that this moves after the ready message, send the routing iinformation of migration Peer in PD-BGP process 1 upper inlet routing table, and the configuration information on PD-BGP process 1 is given PD-BGP process 2; And should move Peer and be labeled as the state that cuts out, the configuration information of migration Peer on PD-BGP process 1 that receives stored among the virtual Peer;
The inlet routing table of migration Peer on PD-BGP process 1 is designated as Adj-RIB-IN11;
230, PD-BGP process 2 stores the routing iinformation that the PD-BGP process 1 that receives sends among the Adj-RIB-IN21 of establishment into, and the routing iinformation that receives is carried out routing;
240, PD-BGP process 2 sends to the CBGP process with the local optimum route that routing obtains;
250, after the CBGP process receives the local optimum route, carry out the routing of global optimum's route;
260, when PD-BGP process 1 routing iinformation of Adj-RIB-IN11 is sent finish after, send route and send and finish message, and send message data that TCP connects corresponding migration Peer and connect to the backup TCP of PD-BGP process 2 to PD-BGP process 2;
270, PD-BGP process 2 sends the message that finishes according to the route that receives, draw the local optimum route and send the notification message that finishes, send this notification message to the CBGP process, its local optimum route of notice CBGP process sends and finishes, and the indication backup TCP connects the reception message;
The receiving course that backup TCP connects message can comprise, and: TCP connects and stops to receive message, then correspondence is moved the message data of having handled of Peer, receives also untreated message data and the message data of also not receiving sends to the backup TCP connection; Perhaps TCP connect to continue receives message, and correspondence is moved the message data of having handled of Peer, received also untreated message data and the message data of also not receiving sends to the backup TCP connection; When the message data of the corresponding migration of TCP connection Peer is identical with the message data of the corresponding migration of backup TCP connection Peer, stop to receive message, then the new connection of backup TCP triggering.
Backup TCP connects to be handled and receives message, such as for original reception but the message that does not have enough time to handle handle, the message that will receive only part originally continues to receive.At this moment backup TCP can response to network.
280, the CBGP process sends the notification message that finishes according to the local optimum route that receives, and selects global optimum's route, sends migration then and finishes message to PD-BGP process 1;
Because it is to select from the routing iinformation of migration Peer that PD-BGP process 2 sends to the local optimum route of CBGP process, PD-BGP process 1 is also selected the local optimum route according to the routing iinformation of migration Peer and is sent to the CBGP process before, therefore the local optimum route of PD-BGP process 1 transmission is identical with the local optimum route that PD-BGP process 2 sends, at this moment the CBGP process need these two local optimum routes of differentiation be from PD-BGP process 1 or from PD-BGP process 2 just, such as distinguishing, perhaps on the local optimum route, mark with process number.
290, PD-BGP process 1 is according to the migration that the receives message that finishes, all routing iinformations of migration Peer among deletion Adj-RIB-IN11 and the Adj-RIB-IN11, the configuration information of deletion migration Peer, and corresponding all message data information of moving Peer in the deletion TCP connection.
CBGP also can delete the local optimum route of the correspondence migration Peer of PD-BGP process 1 transmission then.
When transition process is failed, occur restarting such as PD-BGP process 2, PD-BGP process 1, PD-BGP process 2 and TCP connect the state that returns to before moving so.
The embodiment of the invention can realize Peer dynamic migration between different PD-BGP of BGP and since its not have the configuration repetition, migration velocity is fast, and can guarantee the forwarding do not interrupt not disconnection of session.
As shown in Figure 5, be embodiment of the invention BGP dynamic migration method embodiment three flow charts, present embodiment with BGP Peer from PD-BGP process dynamic migration to the CBGP process, concrete, moving to target CBGP process (present embodiment is designated as CBGP process 3) with migration Peer from source PD-BGP process (present embodiment is designated as PD-BGP process 1) is that example describes, and specifically comprises:
300, after CBGP process 3 receives the dynamic migration order, create virtual Peer and be connected with backup TCP;
The virtual Peer that CBGP process 3 is created is used to store the configuration information that moves Peer; The TCP connection that the backup TCP connection is used to store PD-BGP process 1 is to moving the message data of Peer, and the backup TCP connection did not trigger and set up new connection this moment.
310, CBGP process 3 sends the ready message of migration to PD-BGP process 1, and creates the inlet routing table of migration Peer on CBGP process 3;
The inlet routing table that CBGP process 3 is created is designated as Adj-RIB-IN31;
320, PD-BGP process 1 receives that this moves after the ready message, sends the configuration information of migration Peer on PD-BGP process 1, and gives CBGP process 3 at routing iinformation, the configuration information of PD-BGP process 1 upper inlet routing table; And should move Peer and be labeled as the state that cuts out;
The inlet routing table of migration Peer on PD-BGP process 1 is designated as Adj-RIB-IN11;
330, CBGP process 3 stores the routing iinformation that the PD-BGP process 1 that receives sends among the Adj-RIB-IN31 of establishment into, the configuration information of migration Peer on PD-BGP process 1 that receives stored among the virtual Peer, and the routing iinformation that receives is carried out routing;
340, CBGP process 3 is carried out the routing of global optimum's route according to the local optimum route that routing obtains;
350, when PD-BGP process 1 routing iinformation of Adj-RIB-IN11 is sent finish after, send route and send and finish message, and send message data that TCP connects corresponding migration Peer and connect to the backup TCP of CBGP process 3 to CBGP process 3;
360, CBGP process 3 sends the message that finishes according to the route that receives, and the indication backup TCP connects the reception message;
The receiving course that backup TCP connects message can comprise, and: TCP connects and stops to receive message, then correspondence is moved the message data of having handled of Peer, receives also untreated message data and the message data of also not receiving sends to the backup TCP connection; Perhaps TCP connect to continue receives message, and correspondence is moved the message data of having handled of Peer, received also untreated message data and the message data of also not receiving sends to the backup TCP connection; When the message data of the corresponding migration of TCP connection Peer is identical with the message data of the corresponding migration of backup TCP connection Peer, stop to receive message, then the new connection of backup TCP triggering.
Backup TCP connects to be handled and receives message, such as for original reception but the message that does not have enough time to handle handle, the message that will receive only part originally continues to receive.At this moment backup TCP can response to network.
370, CBGP process 3 sends the message that finishes according to the route that receives, and selects global optimum's route, sends migration then and finishes message to PD-BGP process 1;
Because before migration was carried out, PD-BGP process 1 can be selected the local optimum route and send to CBGP process 3 from the routing iinformation of migration Peer; And the local optimum route that 3 routings of CBGP process draw in this local optimum route and the step 330 is identical, select according to the routing iinformation of migration Peer because all be, therefore CBGP process 3 need these two local optimum routes of differentiation be from PD-BGP process 1 or from CBGP process 3 just, such as distinguishing, perhaps on the local optimum route, mark with process number.
380, PD-BGP process 1 is according to the migration that the receives message that finishes, the routing iinformation of all migration Peer among deletion Adj-RIB-IN11 and the Adj-RIB-IN11, and configuration information, and all data that correspondence is moved Peer are gone up in deletion TCP connections.
CBGP process 3 also can be deleted global optimum's route that its local optimum route that sends according to the PD-BGP process 1 that receives is selected then.
When transition process is failed, break down or restart such as CBGP process 3, PD-BGP process 1 notice TCP connects the migration failure so, and PD-BGP process 1, CBGP process 3 and TCP connect the state that returns to before the migration.
The embodiment of the invention can realize the dynamic migration of Peer between PD-BGP process and CBGP process of BGP and since its not have the configuration repetition, migration velocity is fast, and can guarantee the forwarding do not interrupt not disconnection of session.
As shown in Figure 6, be embodiment of the invention BGP dynamic migration method embodiment four flow charts, present embodiment with BGP Peer from CBGP process dynamic migration to the PD-BGP process, concrete, moving to target P D-BGP process (present embodiment is designated as PD-BGP process 2) with migration Peer from source PD-BGP process (present embodiment is designated as CBGP process 3) is that example describes, and concrete steps comprise:
400, PD-BGP process 2 receives the dynamic migration order, creates virtual Peer and is connected with backup TCP;
The virtual Peer that PD-BGP process 2 is created is used to store the configuration information that moves Peer; Backup TCP connects and is used to store the TCP connection of CBGP process 3 to moving the message data of Peer.
This backup TCP connects not trigger this moment sets up new connection.
410, PD-BGP process 2 sends the ready message of migration to CBGP process 3, and creates the inlet routing table of migration Peer on PD-BGP process 2;
The inlet routing table that PD-BGP process 2 is created is designated as Adj-RIB-IN21;
420, CBGP process 3 receives that this moves after the ready message, send the routing iinformation of migration Peer in CBGP process 3 upper inlet routing tables, and the configuration information on CBGP process 3 is given PD-BGP process 2; And should move Peer and be labeled as the state that cuts out;
The inlet routing table of migration Peer on CBGP process 3 is designated as Adj-RIB-IN31;
430, PD-BGP process 2 stores the routing iinformation that the CBGP process 3 that receives sends among the Adj-RIB-IN21 of establishment into, the configuration information that receives is stored among the virtual Peer and to the routing iinformation that receives carry out routing;
440, PD-BGP process 2 sends to CBGP process 3 with the local optimum route that routing obtains;
450, CBGP process 3 is carried out the routing of global optimum's route according to the local optimum route that routing obtains;
460, when CBGP process 3 routing iinformation of Adj-RIB-IN11 is sent finish after, send route and send and finish message, and send message data that TCP connects corresponding migration Peer and connect to the backup TCP of PD-BGP process 2 to PD-BGP process 2;
470, PD-BGP process 2 sends the message that finishes according to the route that receives, and the indication backup TCP connects the reception message data;
The receiving course that backup TCP connects message data can comprise, and: TCP connects and stops to receive message, then correspondence is moved the message data of having handled of Peer, receives also untreated message data and the message data of also not receiving sends to the backup TCP connection; Perhaps TCP connect to continue receives message, and correspondence is moved the message data of having handled of Peer, received also untreated message data and the message data of also not receiving sends to the backup TCP connection; When the message data of the corresponding migration of TCP connection Peer is identical with the message data of the corresponding migration of backup TCP connection Peer, stop to receive message, the triggering backup TCP connects.
Backup TCP connects to be handled and receives message, such as for original reception but the message that does not have enough time to handle handle, the message that will receive only part originally continues to receive.At this moment backup TCP can response to network.
480, CBGP process 3 with the routing iinformation of Adj-RIB-IN31 send finish after, select global optimum's route;
Because it is to select from the routing iinformation of migration Peer that PD-BGP process 2 sends to the local optimum route of CBGP process 3, CBGP process 3 is selected the local optimum route according to the routing iinformation of migration Peer before, therefore the local optimum route of the local optimum route that sends of PD-BGP process 2 and CBGP process 3 self is identical, at this moment CBGP process 3 just needs to distinguish these two identical local optimum routes, such as distinguishing, perhaps on these two routes, add mark with the process number of these two local optimum routes.
490, all routing iinformations of migration Peer among CBGP process 3 deletion Adj-RIB-IN31 and the Adj-RIB-IN31, migration Peer is at the configuration information of PD-BGP process 2, and all data of deleting corresponding migration Peer in the TCP connection.
When CBGP process 3 also can be deleted and will do not moved Peer and move out then, the local optimum route that the routing iinformation routing of this migration Peer draws.
When transition process is failed, occur restarting or fault such as PD-BGP process 2, PD-BGP process 2, CBGP process 3 and TCP connect the state that returns to before moving so.
The embodiment of the invention can realize the dynamic migration of Peer between CBGP process and PD-BGP process of BGP and since its not have the configuration repetition, migration velocity is fast, and can guarantee the forwarding do not interrupt not disconnection of session.
As shown in Figure 7, be BGP dynamic migration device embodiment one structural representation that the embodiment of the invention provides, present embodiment comprises: migration preparation module 4, route memory module 5 and message receiver module 6.Wherein moving preparation module 4 is used for creating virtual Peer according to the dynamic migration order that receives and is connected with backup TCP, it is placed in the middle that the configuration information of transferring neighbor on the bgp process of source that route memory module 5 is used for obtaining stores the virtual neighboring that moves preparation module 4 establishments into, routing iinformation in the first inlet routing table of transferring neighbor on the bgp process of source that receives is stored in the second inlet routing table of transferring neighbor on the target bgp process of establishment, and carry out routing according to the routing iinformation that stores in the second inlet routing table;
Message receiver module 6 is used for that the message data that the TCP on the bgp process of source connects corresponding transferring neighbor is received backup TCP and connects.
As shown in Figure 8, BGP dynamic migration device embodiment two structural representations that provide for the embodiment of the invention, present embodiment also comprises: local route selection module 8, be used for will store into according to route memory module 5 routing iinformation of the second inlet routing table carry out the center bgp process that local optimum route that routing obtains sends to the routing that is used for global optimum's route.The message sink module 9 that finishes is used for the route that the reception sources bgp process sends and sends the message that finishes.Notification message sending module 10 is used for sending the message that finishes according to the route that the message sink module 9 of finishing sends and draws the local optimum route and send the notification message that finishes, and sends this local optimum route and sends the notification message that finishes to the center bgp process; End sending module 12 is used for local optimum route that the center bgp process sends according to notification message sending module 10 and sends the notification message that finishes and select global optimum's route, and sends the migration end to the source bgp process.Also comprise route removing module 13, be used for the migration end that the source bgp process sends according to the end sending module 12 that receives, the routing iinformation of deletion migration Peer on the bgp process of source.Message removing module 14 is used for the message data that source bgp process notice TCP connects corresponding migration Peer in this TCP connection of deletion.
The embodiment of the invention can realize the dynamic migration of Peer between CBGP process and PD-BGP process of BGP and since its not have the configuration repetition, migration velocity is fast, and can guarantee the forwarding do not interrupt not disconnection of session.It should be noted that at last: above embodiment is only in order to technical scheme of the present invention to be described but not limit it, although the present invention is had been described in detail with reference to preferred embodiment, those of ordinary skill in the art is to be understood that: it still can make amendment or be equal to replacement technical scheme of the present invention, and these modifications or be equal to replacement and also can not make amended technical scheme break away from the spirit and scope of technical solution of the present invention.

Claims (9)

1, a kind of BGP dynamic migration method is characterized in that, comprising:
Receive the dynamic migration order, and create virtual neighbours and back up transmission control protocol TCP according to described dynamic migration order and be connected;
Obtain the routing iinformation of transferring neighbor in the configuration information on the Border Gateway Protocol process of source and the described transferring neighbor first inlet routing table on the Border Gateway Protocol process of source, it is placed in the middle to store described configuration information into described virtual neighboring, the routing iinformation of the described first inlet routing table is stored in the second inlet routing table of transferring neighbor on object boundary gateway protocol process of establishment, and carry out routing according to the routing iinformation that stores in the described second inlet routing table;
The message data that TCP on the Border Gateway Protocol process of described source is connected corresponding described transferring neighbor receives the backup TCP connection.
2, BGP dynamic migration method according to claim 1, it is characterized in that, also comprise: will carry out the local optimum route that routing obtains according to the routing iinformation of the described second inlet routing table and send to center Border Gateway Protocol process, described center Border Gateway Protocol process is carried out routing according to described local optimum route.
3, BGP dynamic migration method according to claim 2 is characterized in that, also comprises:
The route that receives described source Border Gateway Protocol process sends the message that finishes, and sends the message that finishes according to described route and draw the local optimum route and send the notification message that finishes;
Send described local optimum route and send the notification message that finishes to described center Border Gateway Protocol process;
Described center Border Gateway Protocol process sends the notification message that finishes according to described local optimum route and selects global optimum's route, and sends the migration end to described source Border Gateway Protocol process.
4, BGP dynamic migration method according to claim 3 is characterized in that, also comprises: described source Border Gateway Protocol process is deleted the routing iinformation of described transferring neighbor on the Border Gateway Protocol process of described source according to described migration end; Described source Border Gateway Protocol process notifies described TCP to connect the message data of corresponding described transferring neighbor in this TCP connection of deletion.
5, BGP dynamic migration method according to claim 1 is characterized in that, the message data that the TCP on the Border Gateway Protocol process of described source is connected corresponding described transferring neighbor receives the backup TCP connection, specifically comprises:
Described TCP connects and stops to receive message, with the message data of having handled of the described transferring neighbor of correspondence, receive also untreated message data and the message data of also not receiving sends to described backup TCP connection; Perhaps
Described TCP connects and continues to receive message, and with the message data of having handled of the described transferring neighbor of correspondence, receive also untreated message data and the message data of also not receiving sends to described backup TCP connection; The message data that connects corresponding described transferring neighbor as described TCP and described backup TCP connect the message data of corresponding described transferring neighbor when identical, stop to receive message.
6, a kind of BGP dynamic migration device is characterized in that comprising:
The migration preparation module is used for creating virtual neighbours according to the dynamic migration order that receives and is connected with backup TCP;
The route memory module, it is placed in the middle that the configuration information of transferring neighbor on the Border Gateway Protocol process of source that is used for obtaining stores described virtual neighboring into, the transferring neighbor that receives is stored in the second inlet routing table of transferring neighbor on object boundary gateway protocol process of establishment at the routing iinformation of the inlet of first on the Border Gateway Protocol process of source routing table, and carry out routing according to the routing iinformation that stores in the described second inlet routing table;
The message receiver module is used for that the message data that the TCP on the Border Gateway Protocol process of described source connects corresponding described transferring neighbor is received backup TCP and connects.
7, BGP dynamic migration device according to claim 6, it is characterized in that, also comprise: local route selection module is used for the routing iinformation according to the described second inlet routing table is carried out the center Border Gateway Protocol process that local optimum route that routing obtains sends to the routing that is used for global optimum's route.
8, BGP dynamic migration device according to claim 7 is characterized in that, also comprises:
The message sink module that finishes is used to receive the route that described source Border Gateway Protocol process sends and sends the message that finishes;
The notification message sending module is used for sending the message that finishes according to described route and draws the local optimum route and send the notification message that finishes, and sends described local optimum route and sends the notification message that finishes to described center Border Gateway Protocol process;
The end sending module is used for described center Border Gateway Protocol process and selects global optimum's route according to the notification message that described local optimum route transmission finishes, and sends the migration end to described source Border Gateway Protocol process.
9, BGP dynamic migration device according to claim 8 is characterized in that, also comprises:
The route removing module is used for described source Border Gateway Protocol process according to the migration end that receives, and deletes the routing iinformation of described transferring neighbor on the Border Gateway Protocol process of described source;
The message removing module is used for described source Border Gateway Protocol process and notifies described TCP to connect the message data of corresponding described transferring neighbor in this TCP connection of deletion.
CN2008101147347A 2008-06-11 2008-06-11 BGP dynamic transferring method and device Expired - Fee Related CN101605089B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2008101147347A CN101605089B (en) 2008-06-11 2008-06-11 BGP dynamic transferring method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2008101147347A CN101605089B (en) 2008-06-11 2008-06-11 BGP dynamic transferring method and device

Publications (2)

Publication Number Publication Date
CN101605089A true CN101605089A (en) 2009-12-16
CN101605089B CN101605089B (en) 2012-02-22

Family

ID=41470646

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2008101147347A Expired - Fee Related CN101605089B (en) 2008-06-11 2008-06-11 BGP dynamic transferring method and device

Country Status (1)

Country Link
CN (1) CN101605089B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102104540A (en) * 2010-12-30 2011-06-22 杭州华三通信技术有限公司 Method for updating AS (Autonomous System) number and BGP (Border Gateway Protocol) speaker
CN102340410A (en) * 2010-07-21 2012-02-01 中兴通讯股份有限公司 Cluster management system and method
CN110061919A (en) * 2019-04-22 2019-07-26 广西大学 A kind of BGP routing neighbor is established and maintaining method and equipment
CN110858851A (en) * 2018-08-24 2020-03-03 阿里巴巴集团控股有限公司 Data connection restarting method, data processing method and device in live broadcast service
CN112615914A (en) * 2020-12-10 2021-04-06 中通服咨询设计研究院有限公司 Method for transmitting multicast hot standby table entries by using border gateway protocol
CN115208813A (en) * 2022-09-14 2022-10-18 腾讯科技(深圳)有限公司 Method, device, equipment and storage medium for migrating border gateway protocol

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8228818B2 (en) * 2005-06-24 2012-07-24 At&T Intellectual Property Ii, Lp Systems, methods, and devices for monitoring networks
CN100452797C (en) * 2005-07-15 2009-01-14 清华大学 High-available distributed boundary gateway protocol system based on cluster router structure

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102340410A (en) * 2010-07-21 2012-02-01 中兴通讯股份有限公司 Cluster management system and method
CN102340410B (en) * 2010-07-21 2014-09-10 中兴通讯股份有限公司 Cluster management system and method
CN102104540A (en) * 2010-12-30 2011-06-22 杭州华三通信技术有限公司 Method for updating AS (Autonomous System) number and BGP (Border Gateway Protocol) speaker
CN102104540B (en) * 2010-12-30 2014-10-29 杭州华三通信技术有限公司 Method for updating AS (Autonomous System) number and BGP (Border Gateway Protocol) speaker
CN110858851A (en) * 2018-08-24 2020-03-03 阿里巴巴集团控股有限公司 Data connection restarting method, data processing method and device in live broadcast service
CN110061919A (en) * 2019-04-22 2019-07-26 广西大学 A kind of BGP routing neighbor is established and maintaining method and equipment
CN112615914A (en) * 2020-12-10 2021-04-06 中通服咨询设计研究院有限公司 Method for transmitting multicast hot standby table entries by using border gateway protocol
CN112615914B (en) * 2020-12-10 2022-07-05 中国电信股份有限公司江苏分公司 Method for transmitting multicast hot standby table entries by using border gateway protocol
CN115208813A (en) * 2022-09-14 2022-10-18 腾讯科技(深圳)有限公司 Method, device, equipment and storage medium for migrating border gateway protocol
CN115208813B (en) * 2022-09-14 2022-11-25 腾讯科技(深圳)有限公司 Method, device, equipment and storage medium for migrating border gateway protocol

Also Published As

Publication number Publication date
CN101605089B (en) 2012-02-22

Similar Documents

Publication Publication Date Title
EP1832063B1 (en) Virtual multicast routing for a cluster having state synchronization
KR101989333B1 (en) Data forwarding method, device and system in software-defined networking
CN101605089B (en) BGP dynamic transferring method and device
US10104040B2 (en) Wireless local area network WLAN user roaming method, apparatus, and system
EP1853004B1 (en) Access point supporting direct and indirect downstream delivery based on communication characteristics
CN101557616B (en) Method for accessing AP (access point) to switch backup access controller and AP (access point) and AC (access controller)
JP5488979B2 (en) Computer system, controller, switch, and communication method
CN105162704B (en) The method and device of multicast replication in Overlay network
CN101160909B (en) Method and system for implementing virtual router redundancy protocol on a resilient packet ring
JPH04326146A (en) Interconnection method of data processing network and boarder node for implementation thereof
CN102916897B (en) A kind of method and apparatus realizing VRRP load balancing
CN110417665B (en) EVPN networking system and method for multiple Fabric scenes of data center
CN104936254A (en) Wireless network seamless transfer method and device based on software-defined network
CN112615914B (en) Method for transmitting multicast hot standby table entries by using border gateway protocol
WO2014075216A1 (en) Method and network device for establishing virtual cluster
CN111432439B (en) UPF data plane extension and system thereof
EP1418716B1 (en) Communication control system, communication control method, routing controller and router suitably used for the same
JP2007243803A (en) Wireless lan route control system, wireless lan management apparatus, wireless lan access terminal and wireless lan route control method
JP3732745B2 (en) Communication connection establishment method
CN102006301B (en) Method for controlling movement of stream control transmission protocol (SCTP) association among multiple terminals
CN101667968B (en) IPv6 message forwarding method, equipment and system
CN101164299B (en) IP network resource configuring method and IP network
CN110912823B (en) SDN and IPv6 reliable routing method oriented to heterogeneous wired and wireless converged network
KR101021278B1 (en) A Device and method for router control plane redundancy by using OSPF
JP4369882B2 (en) Routing method and network system

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20120222